mlpanda / deeplearning_financial

Attempt to replicate: A deep learning framework for financial time series using stacked autoencoders and long- short term memory

DeepLearning_Financial

Attempt to replicate: A deep learning framework for financial time series using stacked autoencoders and long- short term memory

The original article can be found here: http://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0180944&type=printable

I use the S&P data file provided by the authors here: https://figshare.com/s/acdfb4918c0695405e33

My attempts haven't been succesful so far. Given the very limited comments regarding implementation in the article, it may be the case that I am missing something important, however the results seem too good to be true, so my assuption is that the authors have a bug in their own implementation. I would of course be happy to be proven wrong about this statement ;-)

To run the code:

python run_training.py

This assumes that you have all the packages installed, which I am too lazy to list - python will tell you..

deeplearning_financial's Issues

pretty sure the results in the paper are because WT leaks future info

I tried implementing this too. 99% certain their implementation has the Wavelet transform leaking info. If you just do LSTM on daily close_price your odds are slightly better than a coin toss, adding WT they jump to around 65%. Running WT by windows so you don't leak info probability drops below 50%. SAE isn't going to recover anything from that.

I kinda elaborated a bit here

would love to hear your thoughts

Is the data link invalid?

Hello, I am paying attention to what you are doing, but the data link you provided is https://figshare.com/s/acdfb4918c0695405e33 I can't open it. He doesn't seem to be valid. Can you update the download link?
In addition, I also saw that you provided a piece of data, but can you add a data description? Otherwise I don't know what the meaning of the data is.
Thank you!

Any progress

Hi Marco,
Have you got any update about this replicating this?
I find this work very interesting for the probability to apply the method to nucleosynthesis network calculation.
Look forward to hearing more from you.
Yonglin Zhu
Ph.D. Candidate

When I was running, there was a problem.

RuntimeError: input and target shapes do not match: input [60 x 1], target [60] at c:\users\administrator\downloads\new-builder\win-wheel\pytorch\aten\src\thnn\generic/MSECriterion.c:13